- - -
- - -
Dépôt Institutionnel de l’Université libre de Bruxelles / Université libre de Bruxelles Institutional Repository
Thèse de doctorat/ PhD Thesis Citation APA:
Toussaint, J. (2010). Tumeurs mammaires de grade histologique intermédiaire et ambiguïté biologique: amélioration de l'application clinique du grade tumoral : cancer du sein et grade histologique, mythe ou réalité biologique (Unpublished doctoral dissertation). Université libre de Bruxelles, Faculté de Médecine – Sciences biomédicales, Bruxelles.
Disponible à / Available at permalink : https://dipot.ulb.ac.be/dspace/bitstream/2013/209987/4/55b672c7-70f9-41cd-8cfd-3307016a5207.txt
(English version below)
Cette thèse de doctorat a été numérisée par l’Université libre de Bruxelles. L’auteur qui s’opposerait à sa mise en ligne dans DI-fusion est invité à prendre contact avec l’Université (di-fusion@ulb.be).
Dans le cas où une version électronique native de la thèse existe, l’Université ne peut garantir que la présente version numérisée soit identique à la version électronique native, ni qu’elle soit la version officielle définitive de la thèse.
DI-fusion, le Dépôt Institutionnel de l’Université libre de Bruxelles, recueille la production scientifique de l’Université, mise à disposition en libre accès autant que possible. Les œuvres accessibles dans DI-fusion sont protégées par la législation belge relative aux droits d'auteur et aux droits voisins. Toute personne peut, sans avoir à demander l’autorisation de l’auteur ou de l’ayant-droit, à des fins d’usage privé ou à des fins d’illustration de l’enseignement ou de recherche scientifique, dans la mesure justifiée par le but non lucratif poursuivi, lire, télécharger ou reproduire sur papier ou sur tout autre support, les articles ou des fragments d’autres œuvres, disponibles dans DI-fusion, pour autant que :
Le nom des auteurs, le titre et la référence bibliographique complète soient cités;
L’identifiant unique attribué aux métadonnées dans DI-fusion (permalink) soit indiqué;
Le contenu ne soit pas modifié.
L’œuvre ne peut être stockée dans une autre base de données dans le but d’y donner accès ; l’identifiant unique (permalink) indiqué ci-dessus doit toujours être utilisé pour donner accès à l’œuvre. Toute autre utilisation non mentionnée ci-dessus nécessite l’autorisation de l’auteur de l’œuvre ou de l’ayant droit.
--- English Version ---
This Ph.D. thesis has been digitized by Université libre de Bruxelles. The author who would disagree on its online availability in DI-fusion is invited to contact the University (di-fusion@ulb.be).
If a native electronic version of the thesis exists, the University can guarantee neither that the present digitized version is identical to the native electronic version, nor that it is the definitive official version of the thesis.
DI-fusion is the Institutional Repository of Université libre de Bruxelles; it collects the research output of the University, available on open access as much as possible. The works included in DI-fusion are protected by the Belgian legislation relating to authors’ rights and neighbouring rights.
Any user may, without prior permission from the authors or copyright owners, for private usage or for educational or scientific research purposes, to the extent justified by the non-profit activity, read, download or reproduce on paper or on any other media, the articles or fragments of other works, available in DI-fusion, provided:
The authors, title and full bibliographic details are credited in any copy;
The unique identifier (permalink) for the original metadata page in DI-fusion is indicated;
The content is not changed in any way.
It is not permitted to store the work in another database in order to provide access to it; the unique identifier (permalink) indicated above must always be used to provide access to the work. Any other use not mentioned above requires the authors’ or copyright owners’ permission.
Année académique 2010*2011
UNIVERSITE LIBRE DE BRUXELLES Faculté de Médecine
ULB - Campus Erasme
Bibliothèque des Sciences de la Santé - CP 607 Route de Lennick, 808 (Bât.E)
B- 1070 Bruxelles Tél.: 02/555.61.70
Tumeurs mammaires de grade histologique intermédiaire et ambigüité biologique : amélioration de Inapplication clinique du grade
tumoral
Cancer du sein et grade histologique: mythe ou réalité biologique
Thèse présentée par Jérôme Toussaint
En vue de l’obtention du grade de Docteur en Sciences Bio-médicales
Université Libre de Bruxelles
0034‘P1GS3
Année académique 2010-2011
UNIVERSITE LIBRE DE BRUXELLES Faculté de Médecine
ULB - Campus Erasme
Bibliothèque des Sciences de la Santé - CP 607 Route de Lennick, 808 (Bât.E)
B-1070 Bruxelles Tél.: 02/555.61.70
Tumeurs mammaires de grade histologique intermédiaire et ambigüité biologique : amélioration de Tapplication clinique du grade
tumoral.
Cancer du sein et grade histologique: mythe ou réalité biologique.
Thèse présentée par Jérôme Toussaint
En vue de l'obtention du grade de Docteur en Sciences Bio-médicales
Promoteur de thèse: Professeur Christos Sotiriou
Co-promoteur de thèse: Docteur Virginie Durbecq
Résumé
Les anatomopathologistes disposent d’outils permettant d'assister leurs décisions cliniques et d'évaluer les risques de récidive des patientes atteintes d'un cancer du sein. Parmi ceux-ci, le grade histologique du cancer du sein divise les patientes en trois sous-groupes pour lesquels le grade histologique 1 et 3 sont respectivement associés à de bons et mauvais pronostics. Cependant, cet outil est loin d'être parfait, dû au manque de reproductibilité de ce système et du risque de récurrence intermédiaire, peu informatif, des patients classés dans la catégorie « grade 2 ».
Afin de mieux caractériser ces tumeurs de risque intermédiaire, notre laboratoire a introduit un score appelé « Gene expression Grade Index (GGI) », basé sur l'expression de 97 gènes définis par microarrays. De façon intéressante, ce GGI permet de diviser les patientes de grade histologique 2, sur base de leur profil d'expression, en 2 groupes correspondant aux tumeurs de grade 1 ou aux tumeurs de grade 3. Cependant, bien que le GGI apporte une information importante, son applicabilité clinique est limitée par son prix et la nécessité d'utiliser du matériel congelé.
Durant ce travail de thèse, nous avons transposé la signature microarrays en un test RT-PCR, appelé PCR-GGI, basé sur l'expression de 8 gènes qui permet de reproduire les performances du GGI à partir de tissus congelés ou conservés dans de la paraffine.
Cette amélioration permet de faciliter son utilisation en routine clinique.
De plus, nous avons approfondi notre connaissance du grade histologique, au niveau génomique et transcriptomique, et montré que les tumeurs mammaires [ER-positives) peuvent être divisées en deux groupes : un premier groupe de faible instabilité génomique, exprimant faiblement les gènes de prolifération et présentant un faible risque de récurrence ; et un deuxième groupe de haute instabilité génomique [impliquant principalement des amplifications localisées dans les régions 8q et 20q), une expression importante de gènes de prolifération et un mauvais pronostic.
D'autre part, les carcinomes canalaires in situ (DCIS)
présentant des similarités avec les tumeurs invasives, nous avons
voulu mieux comprendre le comportement du grade tumoral parmi
ces tumeurs pré-invasives. Nous avons donc intégré le PCR-GGI au
VNPI et défini le VNPI-GGI. Comparé au VNPI classique, le VNPI-GGI
identifie mieux les patientes qui vont récidiver tôt dans les groupes de risque intermédiaire et haut, et permet donc d'éviter le sur
traitement.
Cependant, le calcul du VNPI est un travail fastidieux et le PCR- GGI seul ne permet pas de prédire les risques de récidives des DCIS.
Nous avons donc cherché un nouveau marqueur pronostique. Alors, qu'il existe des preuves de plus en plus nombreuses supportant l’importance du rôle anti-tumoral des cellules myoépithéliales, nous avons montré qu'une diminution de l’expression de CDIO - un marqueur des cellules myoépithéliale - était hautement corrélée au risque de récidive. Ces résultats soulignent l'importance tant de l'agressivité de la tumeur que de son environnement directe, dans la progression tumorale.
En terme d'applications, les résultats obtenus durant ce travail
de thèse nous ont permis de développer des outils utilisables par les
cliniciens afin d'améliorer la prise en charge des patientes.
Summary
Traditional histopathological tools routinely used to evaluate breast cancer prognosis are designed to assist physicians in their évaluation of clinical outcome. The histological grade of invasive breast cancer, that assigns patients to one of 3 groups for which histological grade 1 and 3 tumors are respectively associated with lower and higher rate of récurrence, bas long provided clinically important prognostic information. However, this tool is far from perfect due to concern over reproducibility and intermediate risk of récurrence of the histological grade 2 that is not informative for clinical decision.
To better characterize tumors classified as histological grade 2, our group has introduced a score called Gene expression Grade Index (GGI) based on a cassette of 97 genes defined by Microarrays.
Interestingly, the GGI was able to reclassiiy patients with histological grade 2 tumors into 2 groups with distinct clinical outcomes similar to those of histological grade 1 and 3, respectively. However, its clinical applicability still remains expensive and often requires frozen tissue.
During this thesis work, we hâve transposed the GGI onto a qRT-PCR assay, called PCR-GGI, based on a set of 8 genes that could recapitulate in an accurate and reproducible manner the prognostic performance of GGI using both frozen and paraffin-embedded [FFPE) tumor samples, to facilitate its use in clinical practice.
Moreover, we hâve explored histological grade of invasive breast cancer at genomic and transcriptomic level and we hâve shown that two classes of ER-positive invasive breast cancer are observed: a first of low genomic instability, low prolifération gene expression and low risk of récurrence; and a second of high genomic instability [implying a major rôle for amplification of région located on chromosome arms 8q and 20q), high prolifération gene expression and worse prognosis.
In addition, since Ductaî Carcinoma in situ (DCIS) and invasive
breast cancer show concordant biologie behavior, we attempted to
better understand the molecular basis of grade in pre-invasive breast
cancer. We hâve then incorporated the PCR-GGl in the VNPI and
defined the VNPI-GGI to improve its prognostic value. Compared to
the classic VNPI, the VNPI-GGI had a better potential to identify early relapsing patients in the intermediate and high score group, and avoid under treatment in high-risk DCIS patients.
However, VNPI scoring is a tedious work and PCR-GGI alone can't predict récurrence in pre-invasive breast cancer. We aimed then to find news prognosis marker in the field of DCIS. As there is now growing body of evidence supporting the rôle of myoepithelial cells [MECs) as natural tumor suppressors, we bave showed that a decrease of CDIO expression- a surface biomarker of MECs - was significantly associated with an increased risk of relapse.
These results highlight the importance of assessing intrinsic DCIS properties as well as juxta-tumoral stroma, both seems to hâve a major rôle in DCIS progression.
In terms of applications, from these results obtained during
this thesis work, we developed methods applicable into clinical
practice to improve patients' management.
Acknowledgments
I would like to thank those who contributed directly to this thesis:
In the first place, I would like to express my sincere gratitude to Dr.
Christos Sotiriou who bas allowed me to do my thesis in bis lab and who had supervised the project and provided insightful suggestions.
This Project was also co-supervised by Dr. Virginie Durbecq who has followed my work since I arrived in the lab. Her help has been invaluable for my research.
The time is long gone when one scientist could do a research project on his own. Then, I acknowledges for their contributions: Denis Larsimont, Benjamin Haibe-Kains, Christine Desmedt, Ghizlane Rouas, David Brown and Sevilay Altintas, with whom the collaboration has been more than a pleasure. I also thank Carolyn Straehle for her éditorial comments.
For their enthusiasm and the great discussions we hâve every day, I would like to thank ail my colleagues of the Functional Genomics &
Translational Research Unit [Carine Vanderstraeten, Carole Equeter, Naïma Kheddoumi, Jessica Métallo, Marion Maetens, Françoise Rothe, Sandy Haussy, Debora Fumagalli, Samira Majjaj, Michail Ignatiadis, Singhal Sandeep, Sherene Loi, Benjamin Bopp and Françoise Lallemand)
I also extend my thanks to members of my jury who hâve accepted to comment this work, namely Prof. Stéphane Louryan (ULB), Isabelle Salmon (ULB), Philippe Simon (ULB), Pierre Heimann (ULB), Remy Salmon (Institut Curie) and Max Chaffanet (Institut Paoli-Calmettes).
La réalisation d'une thèse repose aussi et surtout sur le soutien de nombreuses personnes qui, souvent sans en comprendre le sujet, contribuent largement à l'aboutissement du travail que représente une telle étude. Je remercie donc tout particulièrement:
Ma famille: Maman, Papa, Valérie, Laurent (et Florian maintenant),
vous êtes tout pour moi.
Agnese qui depuis quelques années a rejoint ce "tout".
Et tous mes amis qui, chacun à leur façon, m’ont apporté un soutien énorme durant ce travail, mais également au long des différentes étapes ma vie. Ils ont toujours été d’une aide incommensurable dans les moments difficiles et je les en remercie.
Merci à tous.
Financial Support
The Work presented in this thesis was supported by the Belgian
Funds for Scientific Research [FNRS) through a FRIA grant.
This thesis has been written under the supervision of Prof.
Christos Sotiriou.
The members of the Jury are:
• Prof. Christos Sotiriou [Institut Jules Bordet, ULBJ;
• Prof. Stéphane Louryan (Hôpital Erasme, ULBJ;
• Prof. Isabelle Salmon [Hôpital Erasme, ULB);
• Prof. Philippe Simon [Hôpital Erasme, ULB);
• Prof. Pierre Heimann [Hôpital Erasme, ULB);
• Prof. Remy Salmon [Institut Curie);
• Prof. Max Chaffanet [Institut Paoli-Calmettes).
Contents
1. Introduction ... 1
1.1. Le Breast Cancer ...1
1.1.1 Incidence of breast cancer 1
1.1.2 Progression of breast cancer 2
1.2. Ductal Canrinoma in situ (DCIS] ... 6
1.2.1 DCIS Classification 6
1.2.2 The Van Nuys Prognostic Index 7
1.2.3 Rôle of Myoepithelial cells 8
1.2.3.1 Tumor suppressor function of myoepithelial cells 9 1.2.3.2 Myoepithelial cells and tumoral progression 10
1.2.3.3 Myoepithelial markers 11
1.3. Invasive breast cancer ... 12
1.3.1 Traditional Approach 12
1.3.2 Histological grade 13
1.3.3 Ki-67: Prolifération markers 14
1.4. Clinical Problems ...15 1.5. Contribution of high-throughput technology in breast cancer research
andtreatment ... 17
1.5.1 Molecular signature in DCIS 17
1.5.2 Molecular signature in invasive breast cancer 18
1.5.2.1 Gene expression profiling 18
1.5.2.2 The Genomic Grade Index 20
1.5.3 Copy number variation in invasive breast cancer 22 1.5.3.1 array Comparative Genomic Hybridization 22
1.5.3.2. aCGH in breast cancer 25
1.5.3.3 Genomic Instability and Breast Cancer 26 1.5.3.4 Genomic alterations associated to the histological
grade 27
2. Justification of the thesis ...28
3. Results 32
Chapter 1: Tumor grade in invasive breast cancer
1. Improvement ofthe clinical applicability ofthe Genomic Grade Index through a qRT-PCR test performed on frozen and formalin-fixed paraffm-embedded tissues.
2. Classification of breast cancer and histological grade 2: a comparative genomic hybridization arraystudy.
Chapter 2: Tumor grade and prognosis marker in DCIS
1. Fine-tuning of the Van Nuys Prognostic Index (VNPI) 2003 by integrating the Genomic Grade Index (GGI): new tools for Ductal Carcinoma In Situ [DCIS].
2. Low CDIO expression identifies high-risk ductal carcinoma in situ (DCIS).
4. Concluding remarks and perspectives 87
5. References 94
6. Annex 113
1. Introduction
Traditionally, breast cancer bas been treated according to histological subtype, extension of diseuse and apparent biological aggressiveness. But with the growing récognition of the large heterogeneity of breast cancer, treatment is becoming increasingly individualized. Since the 1990s, new high throughput technologie such as genomic profiling through microarrays hâve emerged and enabled the study of disease at the molecular level. Today, laboratory-based research and translational aspects of breast cancer hâve improved the ability to treat and cure patients with this disease.
The aim ofthis thesis is to examine histological grade in the area of breast cancer. We focused our research on gene expression profiling in preinvasive and invasive lésions and characterized genomic alterations associated with tumor grade. The objective of this study is to better understand tumor grade and to develop new prédictive tools to improve patient management.
1.1 Breast Cancer
1.1.1 Incidence of breast cancer
What is commonly called "human cancer" in fact comprises
several diseases. Together, they account for about one fifth of ail
death in the industrialized countries. Among them, carcinoma of the
breast is the most frequently diagnosed malignancy in women in the
Western world and the major léthal cancer in European and
American women. According to estimâtes in 2002, there were
1,151,298 new cases of breast cancer diagnosed, 410,712 deaths
caused by breast cancer and more than 44 million women living with
breast cancer worldwide [Veronesi et al. 2005]. In Europe, breast
cancer was by far the most common form of cancer, with an
estimated 370,100 new cases diagnosed in women [27.4% of ail
cancer incident), and the most common form of cancer death, with
129,900 deaths [Boyle et al. 2005]. One out of eight to ten women,
depending on the country, will develop breast cancer during their
lifetime [Ferlay et al. 2001].
1.1.2 Progression of breast cancer
The progression of breast cancer can be characterized by the accumulation of genetic mutations in critical genes accompanied by histological progression from normal epithelium to the development of an invasive breast cancer. In recent décades, a histological model of human breast cancer évolution was prédominant: the classic multi-step model of breast carcinogenesis. The general hypothesis was that some forms of breast carcinoma may arise from established forms of neoplastic lésions that do not hâve the capacity to invade past the basement membrane. This model maintains that tumor progression is a linear process manifesting itself as a sequence of pathologically defined stages in which molecular alterations within normal breast epithelium give rise to atypical hyperplasia [ADH), upon which progressive molecular alterations give rise to ductal carcinoma in situ (DCIS) [Dupont et al. 1985, Lakhani 1999].
Normal Breast Preinvasive lésions Invasive lésion
Figure 1.1: Classic multi-step model of human breast carcinogenesis based on histomorphological and epidemiological data.
Molecular alterations within normal breast epithelium give rise to ADH, upon which progressive molecular alterations give rise to DCIS;
thereafter, additional events occur, resulting in invasive ductal
carcinoma [in blue: épithélial cells, in brown: myoepithelial cells, in
pink: basement membrane]. Immunohistochemistry pictures of normal
breast, preinvasive lésions and an invasive lésion adaptedfrom [Moulis
et al, 2008].
ADH, the first premalignant stage of breast cancer progression, is a proliférative lésion of the breast epithelium, which fulfils some but not ail the criteria of low grade DCIS [Van de Vijver et al. 2003], ADH is small and focal, measuring less than 2 to 3 mm. It is a rare condition, being seen in 4% of symptomatic benign biopsies [Pinder et al. 2003]. The second premalignant stage of breast cancer, DCIS, is defined as a prolifération of malignant épithélial cells within the breast parenchymal structures with no evidence of invasion across the basement membrane [Pinder et al. 2003]. Historically, the pathological distinction between ADH and DCIS is largely morphological and is considered by some to be based upon the size and extent of the épithélial prolifération [Tavassoli et al. 1990].
Additional molecular alterations in DCIS are thought to give rise to the malignant stages of invasive and metastatic carcinoma [Figure 1.1] [Moulis et al. 2008].
Over the past several years, the advance of new high throughput technologies such as genomic profiling to identify expression patterns and copy number variations combining with high spécifie tissue microdissection hâve provided powerful tools to study and re-shape the view of breast cancer progression. Instead of defined breast cancer progression as a single linear pathway, recent molecular genetic evidence supports a "multiple linear pathway model" of progression. The first evidence emerging from several comparative genomic hybridization [CGH] studies suggested that loss of 16q was seen almost exclusively in low and intermediate grade DCIS, while a higher frequency of Iq gain and llq loss was observed in intermediate grade. In contrast, high grade DCIS is characterized by loss of 8p, llq, 13q and 14q and by gain of Iq, 5p, 8q, and 17q and a high level of amplification of 17ql2 and llql3 [Riethdorf et al.
1999].
Comparative gene expression profiling on cDNA microarray
analysis of patient-matched normal versus ADH, normal versus DCIS,
and normal versus invasive breast cancer showed that the major
transcriptome change occurred at the normal to ADH transition, and
that such transcriptional alterations are maintained throughout the
later stages [DCIS and invasive breast cancer) of progression. On a
global level, no consistent major transcriptional changes were
identified between the pre-invasive and invasive stages [Ma et al,
2003]. The different stages of breast cancer progression are
evolutionary products of the same clonal origin and suggest that
genes expressed in the premalignant stages [ADH and DCIS) may reflect the progressive potential of the pathological lésion.
Moreover, distinct gene expression profiles associated with different morphological phenotypes and, in particular, tumor grade, hâve been observed in several studies [Ma et al. 2003, Sotiriou et al.
2006, Desmedt et al. 2006]. Low grade [well differentiated] and high grade (poorly differentiated) breast cancers hâve been demonstrated to possess different gene expression patterns, suggesting that distinct gene expression pathways are associated with the low grade and high grade. This topic is discribed in greater detail in the section depicting invasive breast cancer.
Based upon genetic and gene expression studies, Moulis et al.
suggested a modified model of breast cancer progression, with CGH studies providing evidence of two distinct pathways in the évolution of DCIS and invasive breast cancer [figure 1.2). One pathway is characterized by 16q loss and is observed predominantly in low- grade tumors. The second pathway is characterized by llql3 and 17ql amplification. The gene expression studies showed that breast cancers are stratified along two distinct groups correlating with tumor grade and clinical outcome.
This modified model of breast cancer progression shows
similar biological behaviour in DCIS and invasive breast carcinoma
and highlights a continuum between these two diseases. In line with
this, in the next section we focus on the biological description of DCIS
[section 1.2) first and then invasive breast cancer [section 1.3). These
sections are followed by a discussion of the medical problems posed
by these diseases at the time work on this thesis was begun [section
1.4). This introduction ends with a description of the high-
throughput technology used in breast cancer research and treatment.
Figure 1.2: Modifîed model ofbreast cancer progression: the fîrst
pathway is characterized by genetic alterations that include gain oflq
and loss of 16q that is seen predominantly in low grade ductal
carcinoma in situ (DCIS) and invasive breast cancer. The second
pathway is characterized by amplification ofllqlS and 17ql2 in high-
grade tumors. The divergent two pathways model is provided by gene
expression profîling data generated from ADH, DCIS and invasive
breast cancer. Figure from [Moulis et al. 2008].
1.2 DCIS:
Ductal Carcinoma in situ (DCIS) is essentially characterized by the aberrant prolifération of ductal épithélial cells that do not bave the capacity to invade past the basement membrane [Allred et al.
1994, Schnitt et al. 1997]. As explained in the previous section, there is general consensus that DCIS represents an intermediate step between normal breast tissue and invasive breast cancer, and it is assumed that ail invasive carcinomas of the breast are preceded by DCIS, although this progression will only occur in a portion of patients. Historically, DCIS has been diagnosed in a small proportion of patients presenting with a palpable mass or pathological nipple discharge or, occasionally, as an incidental biopsy finding. But in modem practice, at mammographie screening DCIS is usually detected by typical patterns of microcalcifications [Kessar et al.
2002]. The incidence of carcinoma in situ of the breast accounts for approximately 20% of screen-detected breast cancers, compared to 3-5% of ail symptomatic cancers before the period of population- based mammographie screening [Leonard et al. 2004, Li et al. 2005].
The long-term prognosis of DCIS is excellent, with an expected rate of 10-year overall survival that exceeds 95%, even in the absence of treatment [Boughey et al. 2007]. However, 16-22% of women expérience local relapse within 10 years following lumpectomy alone, and approximately one-half of these relapses are invasive breast cancer. Adjuvant radiotherapy reduces the 10-year risk of relapse to 7-9%, although it is associated with cutaneous toxicity and a small long-term risk of secondary neoplasm [Bijker et al. 2001, Fisher et al. 1998, ]ulien et al. 2000].
1.2.1 DCIS Classification
DCIS represents a heterogeneous spectrum of lésions, varying in morphology, extent and clinical présentation. DCIS is typically classified according to architectural pattern (solid, cribriform, papillary, and micropapillary). These descriptive categories can be of limited clinical utility, particularly because individual DCIS lésions often demonstrate architectural and morphological heterogeneity.
Classification Systems based on tumor grade [high, intermediate, and
low) and the presence or absence of comedo necrosis show the
greatest reproducibility. To improve patient management, modem
classifications separate DCIS into three categories, but differ in the choice of features that are used for categorization. However, there are some features of DCIS, such as tumor size, histological grade, presence of comedo necrosis, margin status, and patient âge that allow some degree of risk stratification for récurrence [Fisher et al.
1998, Silverstein et al. 1999, Jensen et al. 2003].
1.2.2 The Van Nuys Prognostic Index
The Van Nuys classification is the most widely accepted method for risk estimation and has replaced the Holland [based primarily on cytonuclear différentiation and secondary on architectural différentiation or cellular polarization), the Bellamy (based on prédominant architectural pattern as comedo, solid, cribriform micropapillary, or papilary type), the Leal (based on the nuclear morphologie features) and the Lagios (based on nuclear grade, cytoarchitecture and necrosis) classifications [Leong et al.
2001]. The Van Nuys Prognostic Index (VNPI) itself is a simple
scoring method that has been used to stratify patients with different
risks of local récurrence. The index is based upon grade, size,
presence or absence of comedo necrosis and margin width
[Silverstein et al. 1996]. The results from a rétrospective study on the
influence of patient âge hâve led to a modification of the VNPI using
âge as an additional fourth parameter in the scoring System (table
1.1) [chouten van der Velden et al. 2006]. According to the VNPI,
DCIS is classified into three groups with low (sum of the scores = 4 -
6), intermediate (sum of the scores = 7-9), or high (sum of the
scores = 10 - 12) risk of local récurrence after breast conserving
therapy. Three different treatment modalities are advised for the
three different subgroups, respectively, lumpectomy, lumpectomy +
radiotherapy, mastectomy. The simplicity of this tool has enticed
many clinicians to use it as a decision-making aid in the management
of DCIS, but attempts at independent vérification of these proposed
risk stratification schemes hâve not yet consistently validated their
utility.
Score: 1 2 3
Tumor size: (mm) s 15 16-40 > 40
Margin width:
(mm) >10 1 - 10 <1
Grade:
Non high grade no comedo
Non high grade with comedo
High grade with or without
necrosis necrosis comedo necrosis
Age at diagnosis: >60 40-60 <40
Table 1.1: The original Van Nuys Prognostic Index (VNPI) was based upon three predictors for local récurrence: tumor size, margins and pathologicalgrade. Each predictor was given a score from 1 [most favourable) to 3 (most unfavourable]. The sum of the three scores resulted in the VNPI, which varies from 3 to9.A modification was made in the VNPI, introducing âge as a prognostic factor for récurrence leading to the Modified Van Nuys Prognostic Index, which varies from 4 tol2.
1.2.3 Rôle of Myoepithelial cells
The cells that compose the duct of mammary gland are arranged in two layers: the luminal épithélial layer and a more or less continuons layer of myoepithelial cells [MECs) in direct contact with the basement membrane that surround the whole structure.
Functionally, MECs are a hybrid of both smooth muscle and
épithélial cells. Like muscle cells, MECs express filamentous smooth
muscle actin and smooth muscle myosin, and exhibit contractile
properties; like épithélial cells, MECs express intermediate filaments
[the épithélial keratins) [Radnor 1972] and hâve cadherin-mediated
cell-cell jonctions [Franke et al. 1980]. The contractile properties and
the central rôle in milk éjection during lactation are one aspect of
mammary MEC fonction - upon contraction the MECs decrease the
length and increase the diameter of the ducts to eject the milk
[Emerman et al. 1986]. However, during development, MECs also act
to induce luminal cell polarity [Gudjonsson et al. 2002] and to
regulate ductal morphogenesis [Niranjan et al. 1995]. Connection to
basement membrane and desmosomal interactions with luminal
épithélial cells improve paracrine regulatory mechanisms. The MECs appears to be responsible for both the synthesis and dégradation of the mammary gland basement membrane through the production of components such as collagen IV, laminin-1, laminin-5, and fibronectin that regulate ductal growth [Warburton et al. 1982], and through the production of matrix metalloproteinases [Dickson et al. 1992].
1.2.3.1 Tumor suppressor fonction of myoepithelial cells
In past décades the major focus of cancer research has been the transformed épithélial tumor cell itself, while the rôle of myoepithelial cells in tumorigenesis has not been widely explored.
But there is now a growing body of evidence supporting the rôle of MECs as natural tumor suppressors due to their inhibitory effect on varions neoplastic phenotypes, including tumor cell growth, invasion, and angiogenesis [Sternlicht et al. 1997, Deugnier et al. 2002, Lakhani et al. 2001]. MECs also synthesize the basement membrane of the duct and alveoli and form a structural barrier between the luminal épithélial cells and the surrounding stroma, thus physically preventing tumor cell invasion.
Some in vitro coculture and xenograft assays were performed and showed the ability of MECs to inhibit the growth and invasion of breast cancer cells [Nguyen et al. 2000]. These effects hâve been largely attributed to paracrine factors such as ECM proteins, protease inhibitors, and varions growth factors secreted by MECs that exert their effects on the tumor épithélial cells. But some of these factors are still unidentified.
MECs also influence the différentiation and polarity of the adjacent luminal épithélial cells. The polarity is observed in vitro when luminal épithélial cells are cultured in reconstituted basement membrane [matrigel], but lost when the cells are grown in collagen I [Gudjonsson et al. 2002]. However, mixing the luminal épithélial cells with normal MECs was able to restore épithélial cell polarity even in collagen cultures.
Studies using serial analysis of gene expression hâve
demonstrated important différences between DCIS associated MECs
and normal MECs. In particular, compared with normal MECs, DCIS-
associated MECs show down-regulation of a variety of genes involved
in normal fonctions, including oxytocin receptor, laminin, and
thrombospondin, up-regulation of genes for chemokines that
enhance épithélial cell prolifération, migration, and invasion such as
SDF1/CXCL12 and CXCL14, and increased levels of enzymes involved
in the dégradation of extracellular matrix, such as matrix metalloproteinases [Allinen et al. 2004].
1.2.3.2 Myoepithelial cells and tumor progression
The prevailing view of breast tumor progression is tumor épithélial celldriven, since tumor épithélial cells hâve acquired genetic changes and demonstrate genomic instability, and the most aggressive invasive cells can be selected out due to clonal sélection.
But studies showing genetic changes in tumor stroma raise the possibility that clonal sélection occurs in noepithelial cells. Moreover, identification of global gene expression changes and epigenetic alterations in ail cell types during breast tumor progression and the finding that the genetic background of the host influences metastatic behaviour suggest that microenvironment may play an active rôle in driving tumor progression and tumorigenesis may be a "team effort"
[Polyak et al. 2005]. Based on these data, two hypothèses should explain the in situ to invasive carcinoma transition: the "escape” and the "release" models.
Ductal Carcinoma in situ Invasive Carcinoma
»:
Escape model;
Release model:
Figure 1.3: Two alternative models of the in situ to invasive carcinoma transition:
- the escape model: genetic changes in épithélial cells lead to the sélection ofa clone with invasive properties that disrupt the MEC layer, dégradé the basement membrane, and spread into the stroma, and subsequently expand
- the release model: the myoepithelial cells disappear; the basement
membrane is disrupted and results in the release of tumor épithélial
cells. Figure adapted from [Polyak et al. 2005].
In the escape model, genetic changes in épithélial cells lead to the sélection of a clone with invasive properties that disrupt the MEC layer, dégradé the basement membrane, and migrate into the stroma, whereas in the release model, the myoepithelial cells disappear and the basement membrane is disrupted at sites coinciding with areas of leukocytic infiltration and accumulation of myofibroblasts [figure 1.3).
1.2.3.3 Myoepithelial marker
An intact MEC layer is seen in both benign and in situ lésions, whereas loss of the MEC layer is considered as the gold standard for the diagnosis of invasive cancer. It is generally assumed that MEC markers that permit the reliable identification of normal MECs will similarly permit the identification of DCIS-associated MECs [Yaziji et al. 2000]. To differentiate luminal épithélial and myoepithelial cells, and highlight an intact MEC layer, many methods hâve been developed using cell type spécifie markers, many of which hâve been only fortuitously identified following immunohistochemical analysis of breast tissue. Myoepithelial cell spécifie genes include smooth muscle actin [SMA), CDIO/CALLA cell surface marker, calponin, cytokeratins 14 and 17 [CTK14 and CTK17), epidermal growth factor receptor [EGFR), and p63.
Among these markers, CDIO/CALLA [also known as NEP), a
90-100 kDA cell surface zinc-dependent peptidase
[metalloproteinase), is commonly expressed in stromal myoepithelial
cells from normal breast tissue, in 30% of DCIS and is completely lost
in invasive breast cancer [Kalof et al. 2004]. Matrix
metalloproteinases are a family of metallopeptidases that cleave the
protein components of extracellular matrix and thereby play a
central rôle in tissue remodelling.
1.3 Invasive breast cancer
Ductal carcinoma of the breast, also known as infiltrating cancer, occurs when cancerous cells bave spread beyond the ducts of the breast to other parts of the breast.
1.3.1 Traditional approach
The most important goal of invasive breast cancer classifications is to provide a basis for choice of therapy in each individual patient. Surgery is the primary treatment in the majority of cases, alone or in combination with radiotherapy. Despite early détection, up to 50% of women with invasive breast cancer will develop distant metastasis. Metastatic breast cancer is unfortunately incurable. As a resuit, randomized trials of adjuvant systemic therapy bave been conducted in an effort to reduce the rate of récurrence and to prolong the survival of patients with opérable disease. Différences in progression, response to varions therapeutic agents and risk of récurrence bave made decision-making regarding treatment a challenge. Several tools are already available and decisions related to the use of adjuvant systemic thérapies bave relied on traditional clinicopathologic parameters. The risk of récurrence is primarily determined by the âge of the patient, tumor size, histological grade and nodal status. The détermination of expression status of hormonal receptor [estrogen [ER) and progestérone receptors [PgR]), as quantified by immunohistochemistry [IHC), and the expression (IHC) or the gene amplification [fluorescence In situ hybridization, FISH]
status of the HER2 oncogene, also represent good standard practice.
These clinical parameters can provide prognostic information and are summarized in clinical guidelines, such as the National Institute of Health [NIH) [Eifel et al. 2001] in the USA or the St Galien consensus criteria [Goldhirsch et al. 2003] in Europe in order to assist clinicians and patients in adjuvant therapy decision-making.
These variables can also be combined, as they are in the practically useful and validated multivariable outcome prédiction model.
Adjuvant! Online [Olivotto et al. 2005]. This freely available, web- based tool estimâtes the risk of récurrence [or death) with loco
régional therapy alone and with varions systemic adjuvant
treatments, including endocrine therapy and/or chemotherapy.
1.3.2 Histological grade
In the past decade, histological grade has become widely accepted as a powerful indicator of prognosis in breast cancer.
However, criteria for the widely used Scarff-Bloom-Richardson System of grading [Scarff et al. 1968] were critical because concordance between institutions was suboptimally reproducible. Its modification by Elston and Ellis, named the Nottingham combined histologie grade [Elston et al. 1991] is a well-known histo- pathological parameter routinely used in the clinic essentially to describe tumor prolifération and différentiation, which has improved the reproducibility of histological grading. This System is based on a semi-quantitative microscopie évaluation of three morphologie features of tumor cells, including percentage of tubule formation, degree of nuclear pleomorphism, and accurate mitotic count in a defined field area. The sum of these scores subclassifies breast tumors into three groups, as depicted in table 1.2.
Tubule formation :
-Majority of tumor (>75%): 1 -Moderate degree (10-75%): 2 -Little or none (<10%): 3
Histolooical orade 1 : Score 3-5
-Well-differentiated tumor cell,
-Polarized groups of cells that form tubular or duct-like structures.
Mitotic count:
-0-9 Mitoses/10 hpf: 1 -10-19 Mitoses/10 hpf: 2 -20 or > Mitoses/10 hpf: 3
Histolooical orade 2;
Score 6-7
-Moderately differentiated tumor cells,
-Intermediate différentiation stage between well and poorly differentiated tumor cell.
Nuclear pleomorphism : -Small regular uniform cells: 1 -Moderate nuclear size and variation; 2
-Marked Nuclear Variation: 3
Histolooical orade 3:
Score 8-9
-Poorly differentiated tumor cells,
-No tubular structures, -Nuclear pleomorphism
-High mitotic activity. WÊÊ
Table 1.2: Détermination and characteristics of histological grade. Sum ofthe scores of tubule formation, mitotic count, nuclear pleomorphism defïne the histological grade ofa tumor as low (grade
1), intermediate (grade 2) or high (grade 3).
The grade of invasive breast cancer holds key information about malignant behaviour and patient outcome, and is an indicator of disease récurrence and patient death, independently of lymph node status and tumor size [Elston et al. 1991, Schumacher et al.
1993, Lundin et al. 2001]. High-grade tumors resuit in high risk of récurrence, whereas low- and intermediate-grade tumors recur after longer time intervals. The 10-year survival is reduced from 76% for patients with low grade to 39% for those with high-grade tumors.
Untreated patients with low-grade disease hâve a 95% survival rate at 5-years, whereas those with intermediate and high-grade malignancies hâve survival rates at 5 years of 75% and 50%, respectively.
1.3.3 Ki-67: prolifération markers
Markers other than histological grade hâve been used to describe prolifération. However, only Ki-67 présent in ail proliferating cells is currently used in clinical practice and is able to stratify patients into good and poor prognostic categories. It confers a higher risk of relapse and a worse survival in patients with early hreast cancer [de Azambuja et al. 2007]. Increased Ki-67 protein expression correlates with high histological grade, higher mitotic score and estrogen receptor negativity [Tan et al. 2005]. The appropriate cut-off value that distinguishes between high and low proliférative activity in a clinically relevant manner using Ki-67 immunohistochemistry in breast cancer has not been universally established, and cut-off values vary between 10-40% [Spyratos et al.
2002, Offersen et al. 2003], whereas some researchers hâve chosen
mean or médian values. Some authors hâve argued that the choice of
the cut-off point for IHC may dépend on the clinical objective: if Ki-67
is used to exclude patients with slowly proliferating tumors from
chemotherapeutic protocols, a cut-off of 10% will help avoid
overtreatment. In contrast, if Ki-67 is used to identify patients
sensitive to chemotherapy protocols, it is préférable to set the cut-off
at 25% [Spyratos et al. 2002].
1.4 Clinical Problems
There is growing récognition of the large heterogeneity of breast cancer, and classifying patients into broad categories of disease bas facilitated the development of treatment guidelines.
Allocating breast cancer patients based on individual patient and tumor characteristics into smaller subgroups bas become widely accepted practice, and tailoring treatment in this way has emerged as a rational treatment strategy.
Nevertheless despite the fact that an interesting risk stratification index has been found to assess clinical management in DCIS, selecting the appropriate therapeutic approach for individual patients with DCIS remains a major dilemma. The VNPI is a widely accepted method for risk estimation; however, its validity has been questioned [Macausland et al. 2007]. The optimal management of DCIS is therefore not well established, and it is unclear how many women who are diagnosed with DCIS would ever progress to invasive cancer even without treatment. The inability to distinguish indolent from aggressive disease often leads to overly aggressive local treatment for DCIS and, until recently, mastectomy was considered the appropriate treatment for DCIS. The need for intervention is not absolute and the balance of benefit and risk will differ between patients who are symptomatic and those with screen detected pathology. There are now several acceptable treatments for DCIS - including mastectomy, lumpectomy, lumpectomy plus radiotherapy with or without hormonal intervention [Morrow et al.
2002] - yet there is still no uniform consensus about the treatment of DCIS, which results in the large variety of treatment modalities used around the world.
In invasive breast cancer management, even if histological grade has become widely accepted as a powerful indicator of prognosis, the reproducibility of tumor grading is critical because of insurmountable inconsistencies from institution to institution.
Moreover, the modification of the Bloom and Richardson grading
System designed by Elston and Ellis, named the Nottingham
combined histologie grade and based on semi-quantitative évaluation
of morphologie features (percentage of tubule formation, degree of
nuclear pleomorphism, and accurate mitotic count in a defined field
area] hâve made grading criteria quantitative. Unfortunately,
although histologie grade 1 and 3 (with a low or high risk of récurrence, respectively) are clearly differentiated, histologie grade 2 is ambiguous. Indeed, this intermediate group representing a substantial percentage of tumors (30%-60%) is not informative for clinical decision-making because of intermediate risk of récurrence.
In addition, concordance between two pathologists bas been
investigated and found to range from 50% to 85%, with the major
source of inter-observer discrepancy occurring with these
histological grade 2 tumors, making treatment decision for patients
diagnosed with this type of breast cancer a great challenge.
1.5 Contribution of high-throughput technology in breast cancer research and treatment
1.5.1 Molecular signature in DCIS
The histological classification of DCIS has raised much interest over the past decade, and there is general agreement that cytonuclear features play a central rôle in classification. However, it has become clear that the interobserver variation is too substantial to accept histological classification to be used as a solid basis for the treatment of individual patients. The increasing number of gene expression microarray studies represents an important resource in biomédical research. As a resuit, gene expression based diagnosis has entered clinical practice for patient stratification in breast cancer.
However, the intégration and combined analysis of microarray
studies remains still a challenge. To date, there are only a few studies
of gene expression profiling of DCIS hâve been published, and most
hâve focused on the identification of progression-associated genes by
comparison of in situ and invasive disease; moreover,due to the
difficultés in obtaining frozen material from DCIS, these studies are
based on a small number of samples. However, gene expression
profiling studies hâve reported that consistent différences in
expression of a subset of genes can be identified between low-grade
and high-grade DCIS. Seth et al. compared one case of low- to
intermediate grade DCIS with one case of high-grade DCIS with an
invasive component and identified genes upregulated or
downregulated in the low- to intermediate-grade DCIS case [Seth et
al 2003]. Adeyinka and colleagues studied six cases of DCIS with
necrosis and four samples of DCIS without necrosis and identified a
signature of 69 transcripts differentially expressed between these
two groups [Adeyinka et al. 2002]. Ma and colleagues used laser
capture microdissection from paraffin-embedded material followed
by gene expression profiling to identify molecular signatures in
premalignant, preinvasive, and invasive stages of breast cancer. The
results of their study suggested that tumor grade, rather than tumor
stage, is associated with distinct gene expression patterns and that
changes in gene expression required for invasive growth are already
présent in the DCIS stage [Ma et al. 2003]. Hannemann et al. used
supervised classification to identify a gene expression classifier of 43
genes and hâve demonstrated that gene expression profiling can
distinguish between in situ breast cancer samples of well-versus poorly differentiated type. There appear to be a group of poorly differentiated samples, a group of well- and well-intermediately differentiated samples, and a third group containing mainly intermediately-poorly differentiated in situ cases [Hannemann et al.
2006]. Balleine et al. defined a classification tree model including nuclear grade and Ki-67 score, which accurately predicted genomic grade for almost 96%. In an independent cohort of 134 patients witb DCIS, uniformly treated by local excision alone, high molecular grading score was associated witb a pattern of rapid disease récurrence.
The identification of spécifie gene expression patterns in DCIS correlated witb the clinical outcome may help to elucidate the processes underlying the évolution of in situ carcinomas of the breast and also lead improve management of DCIS lésions.
1.5.2 Molecular signature in invasive breast cancer
1.5.2.1 Gene expression profiling
Futher progress in the classification of breast cancer appears to emerge from gene expression profiles, which renders it possible to interrogate tens of thousands of expressed genes simultaneously and to draw up genetic portraits of biological samples. The major objectives of molecular diagnostics are to classify tumors based on molecular characteristics to better understand their biology, evaluate prognosis, and select patients who are at high risk for relapse and therefore could benefit from adjuvant thérapies. Using non- supervised approaches, many early studies hâve show that [1) estrogen receptor [ER] status has the strongest association with gene expression followed by tumor grade; [2] breast tumors can be grouped according to at least four individual subgroups: the basal- like and erbB2 subgroups, which are predominantly ER-negative, and two or more luminal subgroups, which are predominantly ER- positive; and [3] each subgroup has a distinct clinical outcome and may therefore respond differently to varions therapeutics [Pérou et al. 2000, Sorbe et al. 2001, Sorbe et al. 2003, Sotiriou et al. 2003, Hu et al. 2006]. However despite their early promise, classifications generated by hierarchical cluster analysis are extremely unstable.
And although these methods hâve been effective at highlighting
biological différence between tumors, these results hâve not led to
an immédiate clinical application due to their unsupervised approach.
In order to build a more accurate tool for breast cancer prognosis, the Netherlands Cancer Institute first indentified a 70- gene prognostic signature using a set of 78 systemically untreated (node-negative) breast cancer patients whose expression is associated with patients clinical outcome [van't Veer et al. 2002]. The 44 patients classified as low risk (good prognosis group) remained free of distant métastasés at 5 years after diagnosis compared with the 34 patients classified as high risk [poor prognosis group) who had developed distant métastasés within 5 years. The same research group later published a validation study on a larger set of 295 patients, including both node-negative and node-positive breast tumors in treated and untreated patients [van de Vijver et al. 2002].
Later, using a similar approach, Wang et al. identified a 76-gene signature to predict the development of distant métastasés in a sériés of 286 untreated (node-negative) primary breast cancer patients considering ER-positive patients separately from ER-negative patients [Wang et al. 2005]. The same research group provided additional evidence of their 76-gene signature performance in a multi-centric cohort of 180 node-negative untreated breast cancer patients [Foekens et al. 2006].
These two prognostic signatures were validated in a sériés of 302 patients from five different centres in a study conduced by the translational research network founded by the Breast International Group (TRANSBIG) [Buyse et al. 2006, Desmedt et. 2006].
Interestingly, when compared to the two conventional clinical risk classification guidelines [St Galien and NIH consensus) both signatures could outperform them in correctly identifying the low- risk patients, suggesting an improvement in breast cancer management by sparing some women over-treatment and the unnecessary toxicity of chemotherapy.
In 2004, two research groups published a model prédictive of résistance to tamoxifen. Ma et al. conducted a genome-wide analysis of a set of 60 ER-positive patients treated with adjuvant tamoxifen.
The expression signature was reduced to a two-gene expression
ratio, namely HOXB13 and IL17BR, transposed onto a PCR-based
technology using standard formalin-fixed paraffin embedded tissue
[Ma et al. 2005]. Similarly, Paik et al. hâve developed a récurrence
score (RS) based on a panel of 16 cancer-related genes and five
reference genes, which can predict the risk of distant récurrence in
women with node-negative, ER-positive receiving adjuvant
tamoxifen. Corrélation between RS and distant relapse was validated retrospectively in 668 archivai samples of patients with ER-positive, node-negative cancers treated with tamoxifen who were enrolled in the National Surgical Adjuvant Breast and Bowel Project (NSABP) B14 clinical trial [Paik et al. 2004].
1.5.2.2 The Genomic Grade Index
To better understand the molecular basis of histological grade, Sotiriou et al. compared gene expression profiling between high- and low-grade tumors and looked at whether gene expression patterns associated with histological grade could improve prognostic performance, especially within the intermediate grade tumors that display the most heterogeneity in both phenotype and outcome [Sotiriou et al. 2006]. Indeed, some studies provide evidence that high and low grade tumors should be approached as distinct diseases: [1) pathologists hâve always recognized that the different histological grade is associated with different phenotype; (2]
Roylance et al. found distinct genetic différences between low- and high-grade. They observed that the long arm of chromosome 16 appeared to be lost in 65% of low-grade tumors compared with only 16% in high-grade ones, implying that the majority of low-grade tumors do not progress to high-grade tumors during disease progression [Roylance et al. 1999]; [3] finally, low and high grade tumors are associated with different clinical outcome [Elston et al.
1991]. The aims of this study were to examine whether histological grading was associated with distinct gene expression profiles and to explore whether the genetic components associated with these particular cellular States could improve the prognostic value of breast cancer grading. A gene expression cassette of 97 genes differentially expressed between low- and high-grade has been identified in only ER-positive tumors because of the dependence between ER status and histological grade. The majority of these genes was associated with prolifération and cell cycle progression and was overexpressed in high-grade tumors. To summarize the similarity between tumor grade and expression profile, a scoring System called gene expression grade index (GGI) was developed.
Interestingly, when the same gene sélection algorithm was used to
compare the gene expression profiles of histological grade 2 tumors
with the profiles from a group of combined histological grade 1 and 3
tumors, no gene specifically associated with histological grade 2 was
identified. Gene expression profiles of intermediate-grade tumors
looked more like a mixture of low- and high- histological grade, than profiles intermediate between the two. Moreover when examining the prognostic value of these molecular subtypes the GGl was able to reclassify patients with histological grade 2 tumors into two groups with distinct clinical outcomes similar to those of histological grade 1 and 3, respectively.
When the implications of the joint distribution of ER status and GGI in predicting clinical outcome was explored, almost ail ER- negative tumors were associated with a high GGl score, whereas ER- positive tumors were associated with a heterogeneous mixture of gene expression grade index values. The investigation of the association between GGl and relapse-free survival showed that GGl separated ER-positive tumors into low-and high-risk groups. In contrast, among patients with high GGl tumors, ER status was not associated with the risk of récurrence. Therefore, when GGl is known, ER status does not provide additional prognostic information, but when ER status is known, GGl can still improve prognostic accuracy [Sotiriou et al. 2006]. When GGl was compared with the 70- and 76- gene [van't Veer et al. 2002, Wang et al. 2005] signatures in prognostic performance, similar séparation in distant metastasis-free survival between low- and high-risk groups was observed with the three signatures, underscoring the crucial rôle of prolifération genes in breast cancer behaviour and in predicting clinical outcome [Fan et al. 2006]. In a meta-analysis of several breast cancer cohorts, Desmedt et al. hâve confirmed that cellular prolifération was the key biological process and common denominator of different breast cancer signatures (figure 1.4).
Although the prognostic value of different gene signatures
appears to outperform the traditional clinico-pathological
parameters, and the cost of conducting microarray experiments is
decreasing, its clinical applicability still remains expensive and often
requires fresh or snap-frozen tissue. Since paraffin-embedded (FFPE)
tissue from breast tumors are available for every patient, and gene
expression can be measured on FFPE by real-time reverse-
transcriptase-polymerase-chain-reaction (RT-PCR), this RT-PCR
technology may be seen as a valuable, alternative RNA quantitative
method that can be used to move these molecular predictors from
the laboratory to the clinic.
Poor Prognosis
7S-gene~?wgn ature Jvan dtr 7i)ver e? aj.
20021
Waund Signature {Cha^ ta atr-^)4 {
tnvastvtaie^s gtai*
Signaiitre fLiu et ai.
2006}
